English

PILLOW: Enhancing Efficient Instruction Fine-tuning via Prompt Matching

Computation and Language 2024-10-08 v2

Abstract

Instruction fine-tuning has conventionally been employed to adapt Large Language Models (LLMs) to a variety of tasks. Nonetheless, this technique often necessitates substantial computational resources, making it impractical for deployment by individuals or small-scale entities. Recently, Low-Rank Adaptation (LoRA) has become a promising alternative, offering high capabilities on par with full tuning with reduced resource overhead. However, attaining satisfactory performance through the fine-tuning of LoRA is a non-trivial challenge. In this paper, we propose PILLOW, which aims to improve LoRA's performance by a discrimination-based prompting method, leveraging LLMs' In-Context Learning ability. PILLOW incorporates a matching network that selects prompts from a user-defined prompt pool, concatenates the selected prompts with the user instruction as input, and performs inference using the LoRA-fine-tuned LLMs. Trained with Reinforcement Learning, PILLOW exhibits commensurate performance on various evaluation metrics compared with typical instruction fine-tuning methods, utilizing only consumer-grade GPU resources and exhibiting a large reduction in computational costs.

Keywords

Cite

@article{arxiv.2312.05621,
  title  = {PILLOW: Enhancing Efficient Instruction Fine-tuning via Prompt Matching},
  author = {Zhenting Qi and Xiaoyu Tan and Shaojie Shi and Chao Qu and Yinghui Xu and Yuan Qi},
  journal= {arXiv preprint arXiv:2312.05621},
  year   = {2024}
}

Comments

Accepted by EMNLP 2023 (Industry Track), Oral Presentation